drivers/net/wan/hdlc_fr: Add needed_headroom for PVC devices
Subsystems (3)
PATCH v2 1/1 drivers/net/wan/hdlc_fr: Add needed_headroom for PVC devices
PVC devices are virtual devices in this driver stacked on top of the actual HDLC device. They are the devices normal users would use. PVC devices have two types: normal PVC devices and Ethernet-emulating PVC devices.
When transmitting data with PVC devices, the ndo_start_xmit function will prepend a header of 4 or 10 bytes. Currently this driver requests this headroom to be reserved for normal PVC devices by setting their hard_header_len to 10. However, this does not work when these devices are used with AF_PACKET/RAW sockets. Also, this driver does not request this headroom for Ethernet-emulating PVC devices (but deals with this problem by reallocating the skb when needed, which is not optimal).
This patch replaces hard_header_len with needed_headroom, and set needed_headroom for Ethernet-emulating PVC devices, too. This makes the driver to request headroom for all PVC devices in all cases.
---
Change from v1:
English language fix for the commit message.
Changed "Ethernet-emulated" to "Ethernet-emulating" because the device is emulating an Ethernet device, rather than being emulated by an Ethernet device.
I'm sorry for my poor English.
--- drivers/net/wan/hdlc_fr.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/wan/hdlc_fr.c +2 -1
@@ -1041,7 +1041,7 @@ static void pvc_setup(struct net_device *dev) { dev->type =ARPHRD_DLCI ; dev->flags =IFF_POINTOPOINT ; - dev->hard_header_len =10 ; + dev->hard_header_len =0 ; dev->addr_len =2 ;netif_keep_dst (dev); }
@@ -1093,6 +1093,7 @@ static int fr_add_pvc(struct net_device *frad, unsigned int dlci, int type) dev->mtu =HDLC_MAX_MTU ; dev->min_mtu =68 ; dev->max_mtu =HDLC_MAX_MTU ; + dev->needed_headroom =10 ; dev->priv_flags |=IFF_NO_QUEUE ; dev->ml_priv = pvc;
Signature · 1 line
2.25.1
Jakub Kicinski
maintainerQuoted 13 lines · On Wed, 2 Sep 2020 17:06:58 -0700 Xie He wrote:
PVC devices are virtual devices in this driver stacked on top of the actual HDLC device. They are the devices normal users would use. PVC devices have two types: normal PVC devices and Ethernet-emulating PVC devices.
When transmitting data with PVC devices, the
ndo_start_xmitfunction will prepend a header of 4 or 10 bytes. Currently this driver requests this headroom to be reserved for normal PVC devices by setting theirhard_header_lento 10. However, this does not work when these devices are used with AF_PACKET/RAW sockets. Also, this driver does not request this headroom for Ethernet-emulating PVC devices (but deals with this problem by reallocating the skb when needed, which is not optimal).
Quoted 3 lines
This patch replaces
hard_header_lenwithneeded_headroom, and setneeded_headroomfor Ethernet-emulating PVC devices, too. This makes the driver to request headroom for all PVC devices in all cases.
Since this is a tunnel protocol on top of HDLC interfaces, and hdlc_setup_dev() sets dev->hard_header_len = 16; should we actually set the needed_headroom to 10 + 16 = 26? I'm not clear on where/if hdlc devices actually prepend 16 bytes of header, though.
drivers/net/wan/hdlc_fr.c static void pvc_setup(struct net_device *dev) · line 1044
Quoted 10 lines
drivers/net/wan/hdlc_fr.c +1
-1@@ -1041,7 +1041,7 @@ static void pvc_setup(struct net_device *dev) { dev->type =ARPHRD_DLCI ; dev->flags =IFF_POINTOPOINT ; - dev->hard_header_len =10 ; + dev->hard_header_len =0 ;
Is there a need to set this to 0? Will it not be zero after allocation?
Quoted 11 lines
drivers/net/wan/hdlc_fr.c +1
-0dev->addr_len =2 ;netif_keep_dst (dev); }@@ -1093,6 +1093,7 @@ static int fr_add_pvc(struct net_device *frad, unsigned int dlci, int type) dev->mtu =HDLC_MAX_MTU ; dev->min_mtu =68 ; dev->max_mtu =HDLC_MAX_MTU ; + dev->needed_headroom =10 ; dev->priv_flags |=IFF_NO_QUEUE ; dev->ml_priv = pvc;
Thank you for your email, Jakub!
Quoted 5 lines · Jakub Kicinski
Since this is a tunnel protocol on top of HDLC interfaces, and
hdlc_setup_dev()setsdev->hard_header_len= 16; should we actually set theneeded_headroomto 10 + 16 = 26? I'm not clear on where/if hdlc devices actually prepend 16 bytes of header, though.
The HDLC device is not actually prepending any header when it is used with this driver. When the PVC device has prepended its header and handed over the skb to the HDLC device, the HDLC device just hands it over to the hardware driver for transmission without prepending any header.
If we grep "header_ops" and "skb_push" in "hdlc.c" and "hdlc_fr.c", we can see there is no "header_ops" implemented in these two files and all "skb_push" happen in the PVC device in hdlc_fr.c.
For this reason, I have previously submitted a patch to change the value of hard_header_len of the HDLC device from 16 to 0, because it is not actually used.
See:2b7bcd967a0f (drivers/net/wan/hdlc: Change the default of hard_header_len to 0)
Quoted 4 lines
Quoted 4 lines
drivers/net/wan/hdlc_fr.c +0
-0
Quoted 8 lines
drivers/net/wan/hdlc_fr.c static void pvc_setup(struct net_device *dev) · line 1044
Quoted 6 lines
drivers/net/wan/hdlc_fr.c +1
-1@@ -1041,7 +1041,7 @@ static void pvc_setup(struct net_device *dev) { dev->type =ARPHRD_DLCI ; dev->flags =IFF_POINTOPOINT ; - dev->hard_header_len =10 ; + dev->hard_header_len =0 ;Is there a need to set this to 0? Will it not be zero after allocation?
Oh. I understand your point. Theoretically we don't need to set it to 0 because it already has the default value of 0. I'm setting it to 0 only because I want to tell future developers that this value is intentionally set to 0, and it is not carelessly missed out.
Quoted 10 lines · Xie He
The HDLC device is not actually prepending any header when it is used with this driver. When the PVC device has prepended its header and handed over the skb to the HDLC device, the HDLC device just hands it over to the hardware driver for transmission without prepending any header.
If we grep "
header_ops" and "skb_push" in "hdlc.c" and "hdlc_fr.c", we can see there is no "header_ops" implemented in these two files and all "skb_push" happen in the PVC device inhdlc_fr.c.
I want to provide a little more information about the flow after an HDLC device's ndo_start_xmit is called.
An HDLC hardware driver's ndo_start_xmit is required to point to hdlc_start_xmit in hdlc.c. When a HDLC device receives a call to its ndo_start_xmit, hdlc_start_xmit will check if the protocol driver has provided a xmit function. If it has provided this function, hdlc_start_xmit will call it to start transmission. If it has not, hdlc_start_xmit will directly call the hardware driver's function to start transmission. This driver (hdlc_fr) has not provided a xmit function in its hdlc_proto struct, so hdlc_start_xmit will directly call the hardware driver's function to transmit.
So no header will be prepended after ndo_start_xmit is called.
There would not be any header prepended before ndo_start_xmit is called, either, because there is no header_ops implemented in either hdlc.c or hdlc_fr.c.
Quoted 44 lines · Xie He
Thank you for your email, Jakub!
Quoted 5 lines · Jakub Kicinski
Since this is a tunnel protocol on top of HDLC interfaces, and
hdlc_setup_dev()setsdev->hard_header_len= 16; should we actually set theneeded_headroomto 10 + 16 = 26? I'm not clear on where/if hdlc devices actually prepend 16 bytes of header, though.The HDLC device is not actually prepending any header when it is used with this driver. When the PVC device has prepended its header and handed over the skb to the HDLC device, the HDLC device just hands it over to the hardware driver for transmission without prepending any header.
If we grep "
header_ops" and "skb_push" in "hdlc.c" and "hdlc_fr.c", we can see there is no "header_ops" implemented in these two files and all "skb_push" happen in the PVC device inhdlc_fr.c.For this reason, I have previously submitted a patch to change the value of
hard_header_lenof the HDLC device from 16 to 0, because it is not actually used.See:
2b7bcd967a0f(drivers/net/wan/hdlc: Change the default ofhard_header_lento 0)Quoted 4 lines
Quoted 4 lines
drivers/net/wan/hdlc_fr.c +0
-0Quoted 8 lines
drivers/net/wan/hdlc_fr.c static void pvc_setup(struct net_device *dev) · line 1044
Quoted 6 lines
drivers/net/wan/hdlc_fr.c +1
-1@@ -1041,7 +1041,7 @@ static void pvc_setup(struct net_device *dev) { dev->type =ARPHRD_DLCI ; dev->flags =IFF_POINTOPOINT ; - dev->hard_header_len =10 ; + dev->hard_header_len =0 ;Is there a need to set this to 0? Will it not be zero after allocation?
Oh. I understand your point. Theoretically we don't need to set it to 0 because it already has the default value of 0. I'm setting it to 0 only because I want to tell future developers that this value is intentionally set to 0, and it is not carelessly missed out.
Quoted 25 lines · On Fri, 4 Sep 2020 18:57:27 -0700 Xie He wrote:
Quoted 10 lines · Xie He
The HDLC device is not actually prepending any header when it is used with this driver. When the PVC device has prepended its header and handed over the skb to the HDLC device, the HDLC device just hands it over to the hardware driver for transmission without prepending any header.
If we grep "
header_ops" and "skb_push" in "hdlc.c" and "hdlc_fr.c", we can see there is no "header_ops" implemented in these two files and all "skb_push" happen in the PVC device inhdlc_fr.c.I want to provide a little more information about the flow after an HDLC device's
ndo_start_xmitis called.An HDLC hardware driver's
ndo_start_xmitis required to point tohdlc_start_xmitinhdlc.c. When a HDLC device receives a call to itsndo_start_xmit,hdlc_start_xmitwill check if the protocol driver has provided a xmit function. If it has provided this function,hdlc_start_xmitwill call it to start transmission. If it has not,hdlc_start_xmitwill directly call the hardware driver's function to start transmission. This driver (hdlc_fr) has not provided a xmit function in itshdlc_protostruct, sohdlc_start_xmitwill directly call the hardware driver's function to transmit.
Quoted 5 lines
So no header will be prepended after
ndo_start_xmitis called.There would not be any header prepended before
ndo_start_xmitis called, either, because there is noheader_opsimplemented in eitherhdlc.corhdlc_fr.c.
Thank you for the detailed explanation.
Quoted 21 lines · Xie He
Quoted 20 lines
Thank you for your email, Jakub!
Quoted 5 lines · Jakub Kicinski
Since this is a tunnel protocol on top of HDLC interfaces, and
hdlc_setup_dev()setsdev->hard_header_len= 16; should we actually set theneeded_headroomto 10 + 16 = 26? I'm not clear on where/if hdlc devices actually prepend 16 bytes of header, though.The HDLC device is not actually prepending any header when it is used with this driver. When the PVC device has prepended its header and handed over the skb to the HDLC device, the HDLC device just hands it over to the hardware driver for transmission without prepending any header.
If we grep "
header_ops" and "skb_push" in "hdlc.c" and "hdlc_fr.c", we can see there is no "header_ops" implemented in these two files and all "skb_push" happen in the PVC device inhdlc_fr.c.
Quoted 6 lines
Quoted 6 lines
For this reason, I have previously submitted a patch to change the value of
hard_header_lenof the HDLC device from 16 to 0, because it is not actually used.See:
2b7bcd967a0f(drivers/net/wan/hdlc: Change the default ofhard_header_lento 0)
Ah, sorry.. the tree I was looking at did not have this commit.
Quoted 13 lines
Quoted 13 lines
Quoted 12 lines
drivers/net/wan/hdlc_fr.c static void pvc_setup(struct net_device *dev) · line 1044
Quoted 10 lines
drivers/net/wan/hdlc_fr.c +1
-1@@ -1041,7 +1041,7 @@ static void pvc_setup(struct net_device *dev) { dev->type =ARPHRD_DLCI ; dev->flags =IFF_POINTOPOINT ; - dev->hard_header_len =10 ; + dev->hard_header_len =0 ;Is there a need to set this to 0? Will it not be zero after allocation?
Quoted 4 lines
Quoted 4 lines
Oh. I understand your point. Theoretically we don't need to set it to 0 because it already has the default value of 0. I'm setting it to 0 only because I want to tell future developers that this value is intentionally set to 0, and it is not carelessly missed out.
Sounds fair.
Applied to net, thank you!
Quoted 2 lines · Jakub Kicinski
Applied to net, thank you!
Thank you, Jakub!
Quoted 5 lines · Xie He
The HDLC device is not actually prepending any header when it is used with this driver. When the PVC device has prepended its header and handed over the skb to the HDLC device, the HDLC device just hands it over to the hardware driver for transmission without prepending any header.
That's correct. IIRC:
- Cisco and PPP modes add 4 bytes
- Frame Relay adds 4 (specific protocols - mostly IPv4) or 10 (general case) bytes. There is that pvcX->hdlcX transition which adds nothing (the header is already in place when the packet leaves pvcX device).
- Raw mode adds nothing (IPv4 only, though it could be modified for both IPv4/v6 easily)
- Ethernet (hdlc_raw_eth.c) adds normal Ethernet header.
(I had been "unplugged" for some time).
Signature · 5 lines
Krzysztof Halasa Sieć Badawcza Łukasiewicz Przemysłowy Instytut Automatyki i Pomiarów PIAP Al. Jerozolimskie 202, 02-486 Warszawa
Quoted 9 lines · Krzysztof Hałasa
Quoted 5 lines · Xie He
The HDLC device is not actually prepending any header when it is used with this driver. When the PVC device has prepended its header and handed over the skb to the HDLC device, the HDLC device just hands it over to the hardware driver for transmission without prepending any header.
Quoted 8 lines
That's correct. IIRC:
- Cisco and PPP modes add 4 bytes
- Frame Relay adds 4 (specific protocols - mostly IPv4) or 10 (general case) bytes. There is thatpvcX->hdlcXtransition which adds nothing (the header is already in place when the packet leaves pvcX device).
- Raw mode adds nothing (IPv4 only, though it could be modified for both IPv4/v6 easily)
- Ethernet (hdlc_raw_eth.c) adds normal Ethernet header.
Thank you for the nice summary!
I just realized that we can indeed support both IPv4/v6 in hdlc_raw. Both IPv4 and v6 packets have a version field in the header, so we can use it to distinguish v4 and v6 packets on receiving. We can introduce it as a new feature some time.